The hyperbolic taper is popular for benzodiazepine withdrawal, but it is based entirely on pharmacological theory and expert consensus, not clinical trials
Ninety percent of the benzodiazepine tapering guideline rest on expert opinion, not evidence
STUDY: Kleykamp BA et al, J Addict Med 2026;
STUDY TYPE: Systematic review with research recommendations
FUNDING: US Food and Drug Administration (FDA)
Background
Benzodiazepines are recommended short-term, but tapering is difficult, and this review of 57 studies reveals how little we know.
Results
Of 29 guideline recommendations, 26 (90%) were based on expert consensus, not direct data. Most evidence is low quality.
No adjunctive medication stood out. Melatonin, lithium, pregabalin, paroxetine, buspirone, carbamazepine, and others each failed to increase the likelihood of successful taper in controlled studies. While they don’t change the bottom line, some improved limited symptoms, like sleep with pregabalin or melatonin; or anxiety with carbamazepine.
Cognitive behavioral therapy showed modest benefits over usual care in a few studies, but those gains faded within three months.
The studies that did exist were riddled with problems. Tapers moved faster than clinical practice recommends: many cut doses by 25–50% every one to two weeks. No studies evaluated hyperbolic “microtapering” approaches.
Most studies involved women over 50, with benzodiazepine use histories ranging from 2 to more than 20 years. However, older age adults were underrepresented.
What About Ashton?
Working in England, Dr. Heather Ashton developed a manual for benzodiazepine withdrawal. It involves conversion to a long-acting benzo (diazepam), education, and slow individualized taper over months. The manual is based on her experience with approximately 300 patients, but not controlled trials.
In a related condition, alcohol withdrawal, we have a recent randomized trial that compared long-acting (diazepam) with short-acting (lorazepam), and found no difference.
Risks of Tapering
One study found an increase in mortality with tapering (though not randomized), and a survey of 1,200 patients found that 54% reported suicidal ideation during tapering.
Practice Implications
- Careful, personalized care is the call when evidence is lacking. Expert opinions can guide us, but it’s not rigid guidance.
- Despite its limitations, CBT does have the best evidence for benzo WD. Here’s the workbook.
- I start with this brief guide, and follow the symptoms closely with the rating scale attached to it.
— Chris Aiken, MD
Director, Psych Partners
Editor in Chief, Carlat Psychiatry Report
5 comments
A frustrated psych NP
May 30, 2026 at 8:56 am
My take on this is that benzodiazepines should not be used for first-line anxiety management in geriatric patients. However, they are freely prescribed and covered by insurance plans. When they aren’t effective or cause dangerous side effects (which I’ve seen is more often than not), getting people OFF of them is a nightmare that comes with symptoms that leave patients wishing they had never started them. Care staff are n facilities lack proper education to understand PRN usage safety and have no idea how dangerous inconsistent dosing and abrupt cessation can be. There are too many other options to start with benzodiazepines.
pauline dinkelberg
May 31, 2026 at 1:41 am
Unfortunately; also PRN usage can lead to physical dependence.
And may be due to interdose withdrawal patients tend to take them more often.
In the Netherlands whe have a compounding pharmacist who makes (tested; not from crushed tablets, but from the API) lower doses for tapering for almost every benzodiazepine (and Z-drugs!). Which helps many patients to come off.
Switching to diazepam is often not a safe choice. And surely not for the elderly! More sedation, more falls etc.
Sam
May 31, 2026 at 8:33 am
Mixing clinical experience, ASAM and VA/DOD guidelines, I convert to valium equivalent dose for two weeks, then allow patient guided taper with 2.5-5 mg (5-10%) reduction of daily dose every 1-2 weeks. Once we reach 20 mg, maintain at 2 weeks and begin tapering again. The goal is slow and low dose reduction to prevent relapse and promote healthy coping strategies along the way. We must remove the ego and treat the patient safely.
I see good efficacy with prn gabapentin, baclofen, and catapres. Rexulti has been a game changer for primary hypervigilant ptsd patients. There’s a good NEI podcast on the study.
pauline dinkelberg
June 7, 2026 at 7:13 am
However, the ASAM guideline differs (based on input from the best experts—namely, the patients) from the other guidelines; reductions are not made by the same number of milligrams, but by the same percentage of the most recent dose each time.
5 mg from 20 mg = 25% 5 mg from 10 mg = 50%
10% of 20 mg = 2 mg Next step: 10% of 18 mg = 1.8 mg
Diazepam is available in doses as low as 0.1 mg in the Netherlands.
If there is no compounding pharmacist it is important to know: to make liquid suspension you better use milk instead of water.
Chris Aiken, MD
June 7, 2026 at 10:02 am
Good points – another option for liquifying meds is Ora-Plus:
https://chrisaikenmd.com/microdosing